Patient-ventilator asynchrony has been associated with adverse outcomes, and patients with high rates of asynchronies (as defined by an asynchrony index (AI) of greater than 10%) are characterized by longer durations of mechanical ventilation and ICU stay. Objectives: The purpose of this study was to assess the performance of an automatic flow and pressure curve analysis to detect dynamic hyperinflation and specific patterns of asynchrony. Methods: We performed a retrospective analysis on a noninvasive ventilation pressure and flow curves database; from these files, 20 cycles/sequences were blindly selected, after at least 5 minutes of signal stabilization for each patient. Oesophageal pressure was measured in all cases, in order to validate the occurrence of ineffective efforts. Flow curves were independently analyzed by two different experts, who classified them as having or not the different abnormalities that were monitored and assigned an AI for each. Curvex automatically evaluated the same sequences. Curvex is a signal treatment platform, based on the random-distorsion test [1] and multiple mathematical algorithms that automatically detect dynamic hyperinflation (i.e. intrinsic PEEP) [2], and various types of asynchronies: ineffective efforts, short and prolonged inspiration, double and multiple triggering. It allows providing the clinician an overall AI value and the qualification of the different types of asynchrony.